Curriculum

To be eligible for an NSF-funded traineeship in Complex Scene Perception, you must be a U.S. citizen or permanent resident who has been admitted to a Penn Ph.D. program in Psychology, Neuroscience, Computer and Information Science (CIS), or Electrical and Systems Engineering (ESE). As part of your application statement, be sure to indicate that you are interested in the IGERT Complex Scene Perception Program.

Students admitted to a relevant Ph.D. program who are not U.S. citizens or permanent residents may still complete the program in Complex Scene Perception as an Associate, and receive a certificate on their transcript.

Each IGERT trainee completes a Ph.D. in a core discipline (typically Psychology, Computer and Information Science, Electrical and Systems Engineering, or Neuroscience) and also participates in a series of cross-disciplinary research and education experiences that are specific to the Complex Scene Perception IGERT program. Engineering students, for instance, would gain cross-disciplinary experience in Psychology/Neuroscience, with an emphasis on psychophysical research with human (or other animal) subjects. Psychology and Neuroscience students would gain cross-disciplinary experience in Engineering, with an emphasis on computational techniques used in robotics and artificial intelligence. As part of their overall education, IGERT trainees will also learn about teaching, ethics, responsible conduct of research, scientific community building, and career development. The IGERT coursework is compatible with the requirements of each participating graduate group; the details of how each course fits into the group’s requirements vary from group to group.

All students who complete the program are awarded a Certificate, which is noted on their transcript. The requirements for the CSP certificate program are as follows. Further description of the individual program components is below.

  1. Students must complete two foundational courses: “Experimental Methods in Perception” and a course that conveys computational methods for perception scientists. The second requirement is satisfied by several courses offered in the Penn Engineering School.
  2. Students must take one CSP Integration Course. This will vary from year to year and is typically taken after the foundational courses have been completed.
  3. Students must participate in the CSP Seminar (journal club) for at least two years during the course of their graduate training.
  4. The dissertation committee must include members from both core and secondary disciplines, as well as an external member.


Foundational Courses: IGERT trainees are expected to complete the following two courses. At least one of these courses should be completed in the first year of the student’s graduate training. Students who have already done advanced coursework in one of these areas may petition to substitute a more advanced course for one of the two required foundational courses.

Experimental Methods in Perception: This CSP foundational course covers the basic experimental methods and data analysis techniques used in the study of human perception. The course, which is designed to be accessible to graduate students from varied disciplinary backgrounds, combines lectures and laboratory sessions involving the collection and analysis of psychophysical data. Students will learn about the following topics:
* Concept of a psychophysical threshold and how to measure it
* Comparison of adjustment, forced-choice, yes-no, and matching methods
* Theory of signal detection and distinguishing sensitivity changes from criterion shifts
* Stimulus programming
* Display calibration
* Maximum likelihood and related data fitting methods
* Estimation of measurement variability, including bootstrapping methods
* Effective visual representation of experimental data
After completing this course, students will be prepared to conduct original studies of perception. They will have gained experience with several networked experimental rigs and the control software and calibration instrumentation associated with them — all of which will continue to be available to them throughout their CSP program.


Computational Methods for Perception Scientists: Two courses have been approved to satisfy this foundational requirement. They are Machine Perception (CIS 580) and Learning in Robotics (ESE 650). The latter course assumes less engineering background.

Each of these foundational courses will prepare students to take one or more advanced elective courses in that discipline as well. IGERT students are expected to take core graduate courses in the cross-discipline as they progress through the program. The exact courses and their timing should be planned in consultation with and approved by the IGERT steering committee.



Integration Courses

CSP Seminar: In this weekly seminar, CSP faculty and students come together to discuss state-of-the-art publications, distributed across disciplines, using a journal club format. The seminar also provides a forum for CSP trainees to present their own work in progress and to gain experience communicating across disciplinary boundaries. The seminar will be offered each year and participants in the trainee program are encouraged to participate each year. Participation in at least two years is required for the Certificate program.
CSP Integration Course: This required integration course focuses on a particular topic involving a collaborative cross-disciplinary research project, with the topics alternating every year. Students completing the Certificate program are required to take this once.


International Research Opportunity: CSP trainees who have completed their foundational and integration courses are eligible to travel to Tübingen, Germany for a three-month research internship or collaboration through a partnership with the Max-Planck Institute for Biological Cybernetics (http://www.kyb.mpg.de/bu/index.html). Here students have the opportunity to experience a functioning interdisciplinary environment where engineers and psychophysicists work together. They will also become part of the broader international community of perception scientists. Taking advantage of this research opportunity is not a requirement for the Certificate program.



Ph.D. Dissertation Research: All CSP trainees are required to have a five-member dissertation committee, with at least one member each from their core and secondary discipline, including their supervisors, and an external committee member. Committees may be formed as early as the third year, to enable CSP students to plan and complete fully integrated, cross-disciplinary dissertation research.



Credentials: All students who complete the program requirements will be awarded a certificate, which will be acknowledged on their academic transcript, indicating that they have fulfilled the requirements of the Complex Scene Perception graduate training program.